The field of “Trump studies” is vast—but disappointingly homogenous.

One way to understand its paradoxical scope and uniformity is through co-occurrence searches on Google Scholar. A query for “Donald Trump” and “racism,” for instance, would net you more than 8,000 articles published since 2016 in journals or textbooks or cited in academic research. “Donald Trump” and “white supremacy” brings in about 1,800 results. “Donald Trump” and “xenophobia” exceeds 2,400 hits. You get the idea.

These are astonishing numbers considering the slow turnaround time for publication of journal articles, the lag time before an article appears on Google Scholar, the restrictiveness of the search criteria, etc. While there’s certainly some overlap, it is clear that a massive body of literature has rapidly developed around Trump, his supporters, and their alleged racism (one can also find a parallel corpus about Trump and sexism, misogyny, and toxic masculinity).

Survey the titles and abstracts of these works, page after page, and you’ll be hard-pressed to find a single article that challenges the prevailing narrative that Trump voters were driven largely, or even primarily, by anti-minority or white-supremacist sentiments. Given all the ink spilled on this topic by highly educated scholars across a range of disciplines using diverse methodologies and theoretical frameworks, it seems reasonable to assume that the case is airtight: Donald Trump is a racist and so are those who voted for him.  

There’s just one problem: the actual voting data.

• The most decisive votes for Trump in 2016 came from districts that went for Barack Obama in 2008 and 2012. It is unclear why these voters, if horrified at the prospect of a black president, would have voted Obama into office to begin with—let alone given him another four years to advance his agenda when he stood for reelection should they have developed buyers’ remorse.
• The election did not constitute a rejection of Barack Obama: Obama’s popularity remained high throughout the 2016 cycle. In fact, his popularity rose over the course of the campaign, even as the ratings for Trump and Clinton plummeted. More than a year into Trump’s presidency, Obama remains highly popular.
• Trump did not spearhead a white uprising: participation rates among whites were roughly equivalent to those in 2012 and lower than those in 2008. In fact, whites made up a smaller share of the electorate than they had in previous cycles, while Hispanics and Asians were better represented.
• Of those who did turn out, Trump actually won a lower share of the white vote than Mitt Romney. He was able to win because he won a larger share of Hispanics and Asians than his predecessor, along with the largest share of the black vote of any Republican since 2004.

When confronted with these inconvenient facts, the most common retort is something along the lines of “Trump shouldn’t have been a viable candidate to begin with. The fact that he was elected despite his racist, sexist (fill in the blanks) language—that this rhetoric was not disqualifying—suggests that his xenophobia, misogyny, etc. resonate strongly among his supporters.”

But does it really suggest this? A “turnabout test” can be illuminating on this point. Here’s how it works.

Take your interpretation of a given social phenomenon and change the agent or subject, ideally switching actors you are negatively predisposed towards for ones you are positively inclined towards, or vice-versa. If your judgments about the appropriateness of the analysis shift in a substantial way depending on its referent, chances are you are biased and need to rethink things a bit.

For instance:

• Hillary Clinton has a long record of hawkishness, surpassing even most of the Republican candidates in the 2016 cycle. The only other candidate in either party with such an extensive and unqualified record of support for war—any and all—might have been “issue” candidate Lindsay Graham (his issue? Be hawkish!).
• Indeed, during the primaries, Clinton actually relied on the same foreign policy consulting firm that advised hawks Marco Rubio and Ted Cruz. Her general election national security advisor team was populated with many of the same people who oversaw the worst excesses of the Bush II administration.
• Clinton was prominently endorsed or praised (and Trump denounced) by almost every major neoconservative “thought leader” on the block. Some endorsed her early on as the best candidate in either party to realize their agenda. Clinton bragged repeatedly on the campaign trail about her close relationship with Henry Kissinger, whom many on the left view as a war criminal.

Given this clear and unwavering record of hawkishness, from her days as first lady supporting the bombing of Yugoslavia, to her senatorial support for the invasions of Afghanistan and Iraq, to her incessant Obama-era drives for regime change in Libya and Syria, to her campaign trail calls for increased U.S. involvement in Ukraine and doubling down on unqualified support for Israel—should we assert that because her supporters did not find this disqualifying, they must themselves be hawks?

Or to be less charitable: did people vote for Hillary primarily because they support the killing, traumatization and immiseration of others abroad—typically people of color, often of relatively low socio-economic status, often Muslims, etc., in the name of ill-defined “U.S. values and interests”?

No. Many who voted for Clinton condemned her foreign policy at length (or otherwise distanced themselves from it), just as many Republicans denounced Trump’s often irresponsible and racial rhetoric. For most Clinton voters, her hawkishness was an aspect of her record and platform that they found troubling, even abhorrent, but not disqualifying. Maybe there were other issues they cared about more than who would suffer and die in Clinton’s inevitable wars. Maybe they just thought the alternative would be the same, or even worse, in these respects. In any case, it is inappropriate to conflate (often begrudging) tolerance for a key component of a candidate’s record or campaign rhetoric with an endorsement thereof.

But while social researchers appear to take into account these mitigating factors when analyzing Clinton votes, they hardly afford the same to Trump’s. Rather, within many academic and otherwise progressive circles, Trump’s election is held up as an indictment of America’s fundamental racism. 

But is it really?

Looking at the lackluster white turnout in 2016 relative to previous cycles, and Trump’s lower vote share among whites as compared to his predecessor, it could be (it is consistent with the data to argue) that Trump’s rhetoric was a drag on his candidacy among white voters rather than being the key to his electoral success. Indeed, one could plausibly argue that the overwhelming majority of whites did seem to find his rhetoric disqualifying.

Trump only won 37 percent of eligible non-Hispanic whites—meaning the overwhelming majority, 63 percent, did not vote for him. Nearly as many whites, 36 percent, abstained from voting as those who voted for Trump, and an additional 27 percent outright supported other candidates (Clinton, Johnson, Stein, McMullin).

Therefore, even if we assume that 100 percent of those who voted for Trump were motivated primarily by anti-minority or white supremacist sentiments (an assertion that has no empirical basis, to be clear, and that would include millions of minorities as well), we still would not be able to make sweeping claims about whites overall on the basis of such a dubious representative sample. Another turnabout test can help drive this point home.

According to Bureau of Justice statistics, about a third of African-American men can expect to be incarcerated at some point in their lives. On a pie chart, the wedge for “black men expected to do time in the penal system” would look roughly equivalent to the red wedge above.

Would it therefore be appropriate for someone who came across a black guy (like me) to assume criminality on the basis of skin color? Of course not. Yet researchers feel disturbingly confident declaring an epidemic of racism among whites because 37 percent of eligible non-Hispanic white voters cast their ballots for the Republican candidate.

These are just a couple quick examples. I did a deeper dive into how negative priors about Trump and his supporters systematically distort research in the forthcoming volume of The American Sociologist.

Was race a factor in the 2016 election? Of course it was. Race has been a factor in virtually every electoral cycle in U.S. history, including the 2008 and 2012 races that led to the election and reelection of Barack Obama. The specific ways that race influences electoral politics are complex, often subtle and unpredictable. Social researchers generally do understand this—yet they pretty consistently fail to apply that knowledge to the case of Donald Trump and his supporters.

Musa al-Gharbi is a Paul F. Lazarsfeld Fellow in Sociology at Columbia University and a research associate with Heterodox Academy.